The fit indices of the final model were: 2 (524) = 676.8 (p < .001), normed 2 = 1.29; CFI
= .98, RMSEA = .032 (90% confidence interval .024–.039), which indicated good fit to the
observed data. The first and the second of the three lagged paths between psychosocial
conditions and safety climate had significant positive regression weights, whereas the third
lagged path was close to zero (p = .44). This offers some support for hypothesis 1a that
supportive psychosocial conditions have a positive causal effect on safety climate, but the
support was not consistent since this influence decreased over time.s